學術產出-國科會研究計畫

文章檢視/開啟

書目匯出

Google ScholarTM

政大圖書館

引文資訊

TAIR相關學術產出

題名 資料串流上連續型查詢處理技術之研究(I)
其他題名 Research on Continuous Query Processing Techniques over Data Streams(I)
作者 陳良弼
貢獻者 政治大學資訊科學系
行政院國家科學委員會
關鍵詞 資料串流;連續型查詢;頻繁子樹;樣型探勘;查詢規劃;序列資料
日期 2005
上傳時間 12-十一月-2012 11:01:34 (UTC+8)
摘要 由於資料串流(data stream)迥異於傳統資料的特性,再加上眾多新興應用的推波助瀾,最近幾年來資料串流已逐漸成為新興應用中備受矚目的資料型態,舉凡資料串流管理系統(DSMS)的相關理論或應用雛型,皆為當前資料工程領域熱烈討論中的研究課題。本計畫將以研究DSMS 的核心技術⎯連續型查詢(continuous query, CQ)之處理為主軸,發展此核心技術所需的三類關鍵技術:表格資料之連續型查詢處理(RelationalCQ Processing)、查詢與資料串流之監控(Query and Data stream Monitoring),以及序列資料之連續型查詢處理(Sequential CQ Processing)。對應於三類關鍵技術,在本年度計畫執行過程中,我們已完成第一年度預定完成之子技術,分別為可擴充式連續型查詢處理、查詢串流之樣型探勘、多數值串流之內容篩選。在利用查詢串流之樣型探勘支援可擴充式連續型查詢處理的研究中,我們將查詢分視為不同查詢樹,藉由頻繁子樹樣型探勘,找出各查詢間的共用子查詢,並分析單一子查詢的查詢規劃,進而推展出多查詢間可同時使用的全域執行規劃。另外在多數值串流之內容篩選方面,我們利用將查詢小片段分群,並利用滑動視窗將資料切成資料片段,透過資料片段和查詢小片段的群組間相似度計算,和序列資料本身的時序特性,提供刪除機制,加速近似結果的比對過程。
Rapid advances in network commun -ications, software and hardware technologies bring huge amounts of data and form the data as continuous data streams. A data stream is an unbounded sequence of data persistently generated at a high speed. Due to its characters different from data stored in traditional databases and many applications relative to it, an enormous number of researchers pay attention to this research issue. At present, a new type of data management system, called data stream management system (DSMS), has become one of the most popular research areas in data engineering field. One of the kernel technologies in DSMS, namely continuous query processing, is developed in this project. The continuous query processing technology in this project is decomposed into three partitions including relational continuous query processing, query and data stream monitoring and sequential continuous query processing; each of them has been specified in the proposal of this project. In the past one year, we have accomplished the purpose goals of the first year of this project. In the field of relational continuous query processing and query and data stream monitoring, multiple queries can be viewed as query trees, and relying on frequent sub-trees mining, common sub-queries can be found and applied to optimize global execution plan for these multiple queries. In the filed of sequential continuous query processing, we propose a novel method based on n-gram to continuously process queries over event streams to find all approximate answers.
關聯 應用研究
學術補助
研究期間:9408~ 9507
研究經費:823仟元
資料類型 report
dc.contributor 政治大學資訊科學系en_US
dc.contributor 行政院國家科學委員會en_US
dc.creator (作者) 陳良弼zh_TW
dc.date (日期) 2005en_US
dc.date.accessioned 12-十一月-2012 11:01:34 (UTC+8)-
dc.date.available 12-十一月-2012 11:01:34 (UTC+8)-
dc.date.issued (上傳時間) 12-十一月-2012 11:01:34 (UTC+8)-
dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/55365-
dc.description.abstract (摘要) 由於資料串流(data stream)迥異於傳統資料的特性,再加上眾多新興應用的推波助瀾,最近幾年來資料串流已逐漸成為新興應用中備受矚目的資料型態,舉凡資料串流管理系統(DSMS)的相關理論或應用雛型,皆為當前資料工程領域熱烈討論中的研究課題。本計畫將以研究DSMS 的核心技術⎯連續型查詢(continuous query, CQ)之處理為主軸,發展此核心技術所需的三類關鍵技術:表格資料之連續型查詢處理(RelationalCQ Processing)、查詢與資料串流之監控(Query and Data stream Monitoring),以及序列資料之連續型查詢處理(Sequential CQ Processing)。對應於三類關鍵技術,在本年度計畫執行過程中,我們已完成第一年度預定完成之子技術,分別為可擴充式連續型查詢處理、查詢串流之樣型探勘、多數值串流之內容篩選。在利用查詢串流之樣型探勘支援可擴充式連續型查詢處理的研究中,我們將查詢分視為不同查詢樹,藉由頻繁子樹樣型探勘,找出各查詢間的共用子查詢,並分析單一子查詢的查詢規劃,進而推展出多查詢間可同時使用的全域執行規劃。另外在多數值串流之內容篩選方面,我們利用將查詢小片段分群,並利用滑動視窗將資料切成資料片段,透過資料片段和查詢小片段的群組間相似度計算,和序列資料本身的時序特性,提供刪除機制,加速近似結果的比對過程。-
dc.description.abstract (摘要) Rapid advances in network commun -ications, software and hardware technologies bring huge amounts of data and form the data as continuous data streams. A data stream is an unbounded sequence of data persistently generated at a high speed. Due to its characters different from data stored in traditional databases and many applications relative to it, an enormous number of researchers pay attention to this research issue. At present, a new type of data management system, called data stream management system (DSMS), has become one of the most popular research areas in data engineering field. One of the kernel technologies in DSMS, namely continuous query processing, is developed in this project. The continuous query processing technology in this project is decomposed into three partitions including relational continuous query processing, query and data stream monitoring and sequential continuous query processing; each of them has been specified in the proposal of this project. In the past one year, we have accomplished the purpose goals of the first year of this project. In the field of relational continuous query processing and query and data stream monitoring, multiple queries can be viewed as query trees, and relying on frequent sub-trees mining, common sub-queries can be found and applied to optimize global execution plan for these multiple queries. In the filed of sequential continuous query processing, we propose a novel method based on n-gram to continuously process queries over event streams to find all approximate answers.-
dc.language.iso en_US-
dc.relation (關聯) 應用研究en_US
dc.relation (關聯) 學術補助en_US
dc.relation (關聯) 研究期間:9408~ 9507en_US
dc.relation (關聯) 研究經費:823仟元en_US
dc.subject (關鍵詞) 資料串流;連續型查詢;頻繁子樹;樣型探勘;查詢規劃;序列資料en_US
dc.title (題名) 資料串流上連續型查詢處理技術之研究(I)zh_TW
dc.title.alternative (其他題名) Research on Continuous Query Processing Techniques over Data Streams(I)en_US
dc.type (資料類型) reporten